| 摘要: |
| 为提高频谱感知系统在低信噪比环境下对微弱信号的感知性能，提出了一种基于随机共振技术和信息几何理论的频谱感知方法。首先通过随机共振技术增强输入信号的能量，以提高感知信号的信噪比。然后，基于信息几何理论将信号矩阵的协方差矩阵对应成流形上的点，并计算流形上样本点之间的散度距离作为感知信号的特征数据。最后，采用BP神经网络对信号特征数据进行分类，有效避免了决策阈值的计算，快速实现了频谱决策。仿真实验证明，所提方法在低信噪比条件下具有更好的感知性能，有效提高了复杂环境下的频谱检测概率。 |
| 关键词: 认知无线电 协作频谱感知 随机共振 信息几何 BP神经网络 |
| DOI：10.20079/j.issn.1001-893x.220508002 |
|
| 基金项目:国家自然科学基金资助项目（61971147）；广东省研究生教育创新计划项目（2020JGXM040） |
|
| A Cooperative Spectrum Sensing Method Based on Stochastic Resonance and Information Geometry |
| HE Jing,WANG Yonghua,WAN Pin |
| (School of Automation,Guangdong University of Technology,Guangzhou 510006,China) |
| Abstract: |
| To improve the sensing probability of the spectrum sensing system in the low signal-to-noise ratio(SNR) environment,a spectrum sensing method based on stochastic resonance and information geometry(SRIG) is proposed.Firstly,the energy of the input signal is enhanced by the stochastic resonance technology and the SNR of the perceptual signal can be improved.Secondly,according to the IG theory,the covariance matrix of the perceptual signal is corresponded to the points on the manifold,and then the divergence distance between the sample points on the manifold is calculated as the characteristic data of the perceptual signal.Finally,the classification algorithm of deep learning is used to classify the signal characteristic data,which effectively avoids calculating the decision threshold and realizes the spectrum decision quickly.Simulation experiments prove that the proposed method has better sensing performance in low SNR,and effectively improves the probability of spectrum detection in complex environments. |
| Key words: cognitive radio cooperative spectrum sensing stochastic resonance information geometry BP neural network |
《电讯技术》编辑部 